--- title: 'Criei 60 Youtube Shorts Virais em 60 min com Automação IA' source: 'https://youtube.com/watch?v=cE0tBNfZN-s' video_id: 'cE0tBNfZN-s' date: 2026-07-24 duration_sec: 1367 --- # Criei 60 Youtube Shorts Virais em 60 min com Automação IA > Source: [Criei 60 Youtube Shorts Virais em 60 min com Automação IA](https://youtube.com/watch?v=cE0tBNfZN-s) ## Summary This tutorial demonstrates how to build an AI automation that generates viral POV (point-of-view) videos using N8N, GPT-4o, Hugging Face's Flux model, and Runway. The creator walks through setting up a no-code workflow that turns a simple text prompt into a complete POV video in just a few clicks. ### Key Points - **AI fear vs reality** [00:01] — The video opens by contrasting common fears about AI taking jobs with the current reality that AI excels at specific tasks. - **Viral POV accounts** [00:27] — Several TikTok accounts posting POV content have amassed hundreds of thousands of followers and millions of views, e.g., one with 500k followers and 63M views on a single video. - **One-click recreation** [01:08] — The creator claims to have recreated the most viral videos with just one click, setting up the need for full automation. - **Automation goal** [01:22] — The goal is to create an AI automation that generates any POV video with a single click, using no-code tools. - **Tools overview** [02:01] — Tools used: N8N (free), ChatGPT (free), Hugging Face (free), Runway ($0.25 per video). Links are provided in the description. - **Teaching approach** [02:31] — The creator emphasizes teaching how to build the automation rather than just providing a ready-made solution. - **Building from scratch** [03:13] — The tutorial starts from scratch in N8N: create a workflow, add a form with a title and user input field. - **English-only fields** [04:06] — Fields must be in English because the AI tools are optimized for English commands. - **OpenAI API setup** [05:47] — Create an API key from OpenAI, paste it into N8N, and select the GPT-4o model. - **GPT prompt** [06:43] — A custom system prompt is used to transform user input into a detailed POV prompt for image generation. - **Image generation HTTP request** [08:09] — Add an HTTP POST request to Hugging Face's Flux model for image generation. Requires a Hugging Face token. - **Why Flux** [10:08] — Flux is chosen because it's free and optimized for high-quality POV images. - **Extract image data** [12:09] — Use an extract node to convert the image response into a JSON format that can be passed to the video generation step. - **Video generation with Runway** [13:00] — Add another HTTP POST request to Runway's API to generate the video from the image prompt. Costs $0.25 per video. - **Wait node** [15:29] — Add a wait node of 40 seconds to allow Runway to process the video. - **Get video link** [15:43] — A third HTTP request retrieves the finished video URL from Runway. - **Output options** [17:38] — The automation can output to Google Sheets (new row) or send the link via Gmail. - **Testing** [20:30] — The creator tests with prompts like 'You are a slave in the construction of the pyramids in Egypt' and verifies the results in the spreadsheet. - **Automation complete** [21:52] — The automation is fully functional: generates prompt, image, video, and delivers the link. - **Future plans** [22:07] — The creator hints at a follow-up video for longer POV videos if this video reaches 100k views in a week. ### Conclusion This tutorial provides a complete no-code automation for generating viral POV videos using AI, demonstrating that such content creation can be both accessible and scalable. ## Transcript today, like I used to, are looking at artificial intelligence as something that will take everyone's jobs. The problem with artificial intelligence today is that we can make it reach an extremely frightening level of competence in a very specific skill. I understood. What a person can do nowadays with this tool is so incredible, man, and it's so revolutionary . So, that TikTok account has almost 500,000 of my followers posting almost 500,000 of my followers posting only one type of content: private videos. already reached over 20 million views. And there's this one, for example, which reached 63 million views. Another viral account that only posts POV videos is this one, which already has over 300,000 followers. I recently came across a POV video on TikTok with over 40 million views, posted by the account you're watching now. Another recently created account posts POV content exclusively for Brazilians and already has over 100,000 followers. What you might not realize is that creating this type of content is much easier than it seems. I even recreated the most viral videos from those accounts with just one click. They are appearing on your screen right now. something much better than creating all those viral videos with you. I'm going to create an artificial intelligence automation that generates any POV video that generates any POV video you want with a single click. But there's a catch. In another class, I also created an automation that generated long card videos. I created a 3-minute video for class, but I had to people who watched this video said: "Daniel, it's very difficult to write code, even if the Portuguese chat writes it for me. Is there a way to automate all of this without code?" It exists. In this video, I've decided that I'm going to automate this process without needing any code. And to begin, let's look at the video tools. First tool, the creation environment. It's going to be N8N, which is free. The artificial intelligence tools we're going to use in this video are GPT chat, which is free, Rugin Face, which also doesn't have any cost, and Runaway, which will charge you 25 cents per video, okay? In the description of this video, I've already organized and detailed all the links and everything you'll need to replicate this. I also did the favor of compiling all my work, downloading it, and leaving a file there. Ready. If you artificial intelligence I created in this video will be ready for you with just one click. But better than handing you something ready-made is teaching you how to do it, because then you learn to do whatever you want. Let's begin. So, since I want you to learn how to create your own teach you by doing it on my computer. First step, come here, to n8.iu. If you don't have an account, click here to get started and create your account. It's free. And there's also a free plan. If you already have an account like me, you'll click on login. I 'm recording this video with 3% battery left, but creating this is so quick that I'll finish everything before my PC battery dies. Pay close attention, then, because I'm going to build everything from scratch with you. There will be an "too many" icon in the upper right corner. You will click here on workflow. You can choose any one. I'm going to choose My Project. OK. My project, when it starts, will have an extra symbol right in the middle of your screen. You click here, on the fifth, fifth request, to create a submission. He will create a test link. Authentication is not required. Form title. Let's write a title like title like gereids p. Write down an order, an gereids p. Write down an order, an idea. A simple and straightforward idea, okay? This is what will appear to the user when they enter your artificial intelligence, your automation. From this point on, pay close attention, okay? Because from now on, the fields will be filled in English. Why? Because English is the native language of most, if not all, artificial intelligences that exist. And since I'm creating a method for you to make these videos practically for free, I need to use IASs that only understand commands in English. So here in the name field I'm going to English. So here in the name field I'm going to write "user input," which would be like a user command. I'll keep this text here in order. And the order is the command that will be responsible for creating your text, okay? Activate this key here. It's place order. I can even write in Portuguese. Yes, write your idea in English. Test. Okay, he generated the request. Write your idea in English. Let's use Google Translate here. So, you So, you woke up in the royal palace of Egypt, or woke up in the royal palace of Egypt, or you are a queen in the royal palace of Egypt. Let's copy this and use this idea as the prompt. OK. pin" because we'll keep this prompt here for our future integrations. Exit, there will be a " more" tab, click on that tab and we will add Openi. So you search for Open here, message template. What are is your credentials. I already have my credentials registered, but if you didn't , you would come here to create new credentials. That's what you're going to do. Once you have the credentials to fill out, click on this link here, Open APIs. You will be redirected to your account on Open APIs, log in to your account on Open APIs, log in to your account, and register the pending API . Wonder. These are the API keys that I already have created to make requests. Come here, create a new secret key. Yes, type, for example, "po videos" and create. Since I already have one, I'll use the one that's ready. It generated the key, copy it. Paste it here in API K. Organization ID is not required. The base URL will be automatically filled in . If it's not there, just leave it here. Once you've entered the API key, it will show "connection completed it will show "connection completed successfully." Text maintains and messages. OK. The model we're going to use. I'm going to select the standard template we use for everything, which is the GPT 40 or 40 chat. Once you've selected the GPT chat template, go here to the YouTube description. I'm going to save this command in the YouTube description. I'm going to save this command in the YouTube description. And it copies the GPT order. Daniel, what will this GPT order do? Basically, it will inform the GPT chat what guidelines it needs to follow to transform your command into a POV video generator prompt worked. Running. It worked. Stage two, functional. This is the first part of our video that's ready. What is the first part of our video? The generation of prompts. So, let's rename this so we can organize everything. Rename requires a submit prompt. I'm not fluent enough in English to write everything here perfectly, okay? But I will write to the best of my ability. It's openi, it's generate prompt. I think that's how you spell it. I'm not going to use a translator to translate everything. And I'm going to generate a note here. You can generate the grades in Portuguese, okay? That they are a representation of the finished work. This representation of the finished work. This note here is the note here is the generation generating the prompt. OK. We're done with this one, we're moving on to the next request. And what is the request that generates the images? For this, we will use an HTTP request. We will select the POST request template here. The example URL is the one you'll use, which is also in the description. Which URL is in the description, which is the example URL? This one here, to generate the image. Daniel, what link is this? I don't understand anything. This link here is for Hugin Face or Ruging Face, I don't know how to pronounce that. I know how to use it. What is this decided to host artificial intelligences that are far superior to the ones you know for free, and which the general public doesn't have access to, all in one place, like a forum called Hugging Face, Hugging Face, I don't care, I use it and I'm good Daniel, how am I supposed to access this from here? So you're going to create an account on Ruging Face. If you don't have an account, click here to create your account. If you verify that I'm human, because I already have an account. I'm going to come here, to log in. Once you've created your account here, you'll click on the link that's also in the YouTube description. What link is this? Create tokens. Let's go to create tokens. Clicked the link to create a new token. Write, which is a token that new token. Write, which is a token that does everything. POV videos, okay? But I already have it, so I'll keep the same token I have, which I created specifically to make this AI here. With the token copied, you'll the token copied, you'll go back to the link in the description and then click on go back to the link in the description and then click on this link here. Flux will be the artificial intelligence responsible for creating our POV images. Will it be Flux and not some other artificial intelligence? Because first of all, it's free. Secondly, this is a version developed to optimize and generate POV images with better quality than any other generating model. So this is the one we're going to use. Understood? Wonder. So, going back to Understood? Wonder. So, going back to our N8N post-processing method. Daniel, which link should our N8N post-processing method. Daniel, which link should I put here in the URL field? The I put here in the URL field? The URLs are in the description below. API URLs are in the description below. API reference cont C conttrl V. OK. You're going to activate the key that's located at the very end and set up generic authentication. Customized authentication. Here, look. Can you see that I have the Hugin Face created? You will create a new credential and you will copy your API key. How do I copy my API key? You're going to copy this my API key? You're going to copy this code here, paste it into N8N, and code here, paste it into N8N, and replace your API key in this place replace your API key in this place here, for example, in the API key that I here, for example, in the API key that I created and left ready here. Let's edit it. I replaced the code with the API key. I'll have to delete this API key after the video, but that's okay. OK. Once you've entered that code exactly as I did in the video description and replaced your API key, your image generator should be working. On this bar, you will activate the last key bar here in your request. Jason's file will be pre-configured. You will select the first option here, name. Remember that I'm going to put it in English? So I'm going to write user input. that I wrote write user input. that I wrote useless. Impils. And here in the value field, you useless. Impils. And here in the value field, you will drag the GPT chat code here. Wonder. Let's test if it's working. Playing the note. Image working. Playing the note. Image generated. Wonder. Then you saw the image in POV. And now, Daniel? Now you can close this process because this integration step is working, and we need to understand that the most efficient and highest quality way to generate a video is by transforming the image into code. Then transforming the image into code. Then add an extract so you can extract the file. Select that last option here. And here you don't need to do anything. Why, Daniel? Because the way it comes by default, it's already configured to transform the image into code, a Jon-like code that you won't understand at all. I'll even click on "test" so you can see what's generated here. This code here is generated, and it's huge, and you don't understand it. But for the this is much better than attaching the image. Wonder. Code completed. The second stage of our video is ready, which is generating the image. Now we're moving on to the third stage. So what are we going to do now? Generate the video. Let's add one more HTTP request. Similarly, post here as a URL; you'll copy the URL that's in the description, which is this one here. Copy, come back here and is this one here. Copy, come back here and paste. Customized generic authentication. Can you see that I already have a "han wait" on two registered accounts? So it is. How is it registered? You're going to come here, copy the code that's in the description, which is this code here, press Ctrl+C, paste the code into Jason Ctrl+V, and replace this step here with your API key. And how do I get my API key? Go to the my API key? Go to the YouTube description; you'll find the Hanawayi API generator YouTube description; you'll find the Hanawayi API generator . This is the one you're going to use. created, see? Wonder. You will create a new API key. I'll put POV videos here, for example . To create. Here's a key that I this video is over. And you'll take a copy of it, come here, and paste it into this step here. I already copied it, so here at the end you will activate this Bowe request. Okay. So what are we going to fill in here? You're going to come here and use Jason. You're going to copy and paste the use Jason. You're going to copy and paste the second R code into it, which is this code second R code into it, which is this code here that I saved as bour. OK. Control C, Control V. Okay? You noticed that there was a small error here, right? Why did he make a small mistake? Because my goal is to move the GPT chat text prompt here. Since the nomenclatures are not yet standardized, it should be listed as HTT. Yeah, he's showing as HTT because he hasn't returned, it turned green, okay? HTT because he hasn't returned, it turned green, okay? But to avoid this becoming a problem, let's rename this step in the videos, okay? Let me just check to see if it's running correctly by running your node. That's probably right, everything is ready. Oh, ID executed successfully. It's done. OK. I'll rename it to something like create video, texture video. OK. The runaway takes about 40 seconds to generate. So we're going to add a white and we're going to put 40 seconds here because that's the 30 to 40 seconds it will take to generate the data. Finally, we'll add the last we'll add the last HTTP to make a link request. We're going to ask him to provide the link to the video generated by Hanaway. The URL prepared. You won't have to do anything at all. Here, press Ctrl+V. You'll activate it as usual. Ah, it's different here. Here you will activate the second parameter, the second gear switch. To use. Here you will select the first part. In authorization. Control C. Go back here. Control V. Value. Valor is your API key. Remember the one you copied? You can copy it again. Either you're a fool, or you didn't save her. We'll be back here, okay? Let's open. Let's edit to view our API key. We're going to copy this and paste it into our going to copy this and paste it into our request. Open, which is the request for value request. Open, which is the request for value here. Ready. Let's add one more parameter here, okay? The second parameter we're going to add is the type, which is this one here. Hanaway's version. Ctrl+C comes here. Ctrl+V. Ctrl+C comes here. Ctrl+V. This value is here. Control Ctrl V. Wonderful. Let's test if it will pull the request correctly. There was an this because I made a mistake copying and pasting quickly, because my battery is dying. I didn't copy this step here, look. Yes, you have to copy this step here, look. Yes, you have to copy this request here as well, otherwise it will definitely give an error. Now let's test it again and it will probably work . Video successfully generated. This is the video he made. Video successfully generated, but it's still a test. Once the video is ready, for it to be 100% automated, you need to receive the link automatically. So now all you have to do is to be sent. Ah, Daniel, I want the link to be sent to a Google sheet, Google Sheet. The second one is to create a new line. It will connect your account. Mine is already connected, but if it was n't, I could just press this button one line. No, a spreadsheet. I think that must be it, it's written. Yes, OK. Add documents from your list. I created a spreadsheet, POV videos. From your list, he will pull the pages. Mine only has one page. It will map the columns automatically. And you're going to put it here, ID status, man. It's put it here, ID status, man. It's Okay, it's integrated. He will add it to the spreadsheet. Oh, Daniel, I want him to send it to me by email. So you're going to come here and you're going to create a new Gmail node. You will connect your Google account. Why? For him to send me a Google account. Why? For him to send me a message. Gmail. OK. Oh, just press this button and the account is connected. Mine has already gone through that process. Who are you that process. Who are you going to send the email to? For myself. Danielgonzagatj@gmail.com. It's a subject and a product that has been successfully generated. OK. and a product that has been successfully generated. OK. Email has HTML in the email. Message: "Here's the link to your production." I'll just post this one; I won't even include a POV because that's faster. You come here in expressions like that. Ah, since he's not connected, he won't be able to retrieve the printouts. So, let retrieve the printouts. So, let me make a connection here from our Gmail me make a connection here from our Gmail to our request. Okay, now it 's sending the requests. Since the requests are here, I think I can pull them using fixed. Here's the link to your production. OK, saved. Ready. Now, if you want to do the can already do it, it's already working. But I'll organize this by notes. The first part we did was generate the prompt. The second part we did was generate the image. Okay, let me rename it by generating your P.O.V. image. Done. I'll change the color too. I'm going to put on a red one. I'm enlarging my screen here to make it easier to navigate. OK. Up to this point, it has shaped his image. After this point, we're in the video generation phase. So, one more note. Add a note when creating your one more note. Add a note when creating your videos. Creating a videos. Creating a green P, join it here, extend it until the green P, join it here, extend it until the complete creation, which is this HTT. And in the end it's clear that he's going to send the generated video to a spreadsheet or to a Gmail account. The best way to test this is live, right? So let's think about the prompts that work. A viral prompt, including the line "you are a slave in the construction of the pyramids in Egypt," I'll copy this from here, I'll come to artificial intelligence to test it. She's going to open a link. Write your idea in English. Submit. Go back to her. Generating prompt. He's going to send an OK message. OK. Prompt generated. Generating image. OK. Image generated. File extracted. Generating POV. Request here on Hunaway, running the POV. Request here on Hunaway, running the video request. Tam. Ok, video generated. Please wait 40 seconds while you extract the video. Time's up. He added it to Sheets and sent a message to my email. Let's check it out here. There you go. Production successfully generated. This spreadsheet contains generated videos. So we asked: "You are a slave in the pyramids of Egypt." He replied: "You are a slave in the pyramids of Egypt." Wonder. We ordered another one. Hey, are you out of touch with reality? Are you daydreaming? You're on Mars. He spawned the one from Mars. Wonder. We ordered another one. This one must be the one from the moon here. You're daydreaming. Wonder. You're on the space station. Wonder. Artificial intelligence created. Automation completed successfully. Here I'm going to download the AI ​​that I created. I'll upload her template in the description of this video. I'll also leave the full description here for you. Daniel, but I saw that some POVs... I also have to add several scenes from the same environment, and I also have to generate the audio, the sound effects, the music. So, how did I create this ? 15, 20 minutes or less, right? Much less. To help you create a system that generates increasingly longer videos. If you enjoyed this video, you'll like and share it. If this video reaches 100,000 views in a week, I guarantee that the next video will be about creating an artificial intelligence automation that generates long that generates long POV videos. Thank you for your attention. Thank you very much.